It is demonstrated, by a series of marker experiments in which layers of nonporous anodic oxide were tagged by incorporating radioactive material in them, that nonporous anodic oxide grows at, or close to, the oxide ‐electrolyte interface; it is concluded that the aluminum ion is the mobile species. Similar experiments with a porous oxide may also be interpreted in terms of formation of the oxide at the oxide ‐electrolyte interface, but in this case the interface is at the bottom of pores filled with electrolyte.